Adsourcer Developer
Publisher integration
Add sponsored recommendations to AI applications without surrendering control of your UI.
Markdown for agents · llms-full.txt
Integration checklist
- Obtain a publisher API key
- Install the SDK or call REST directly
- Send user queries to
POST /v1/ads/recommend - Handle
SHOWandNO_ADdecisions - Render ads with mandatory disclosure labels
- Link CTAs to
click_url - Optionally record impression events on render
Request shape
json
{ "query": "best running shoes under €150", "context": { "currency": "EUR", "location": "NO" }, "session_id": "anon-session-uuid", "placement": "recommendation", "max_results": 1}Rendering ads
Responses include structured fields — title, price, image URL, disclosure — but no HTML. You control layout, typography, and placement. Intent gating ensures ads only appear when commercial intent is detected.
Required disclosure
Display thedisclosure field (typically "Sponsored") adjacent to every ad unit. This is a platform requirement, not optional styling.Prefer a drop-in unit? See Embed for React components and iframe embeds.
Placements
Pass a placement hint when calling recommend — e.g. recommendation, inline, or sidebar. Placements inform ranking and reporting.
Next steps
Questions? See the help center.